Dense depth estimation from multiple 360-degree images using virtual depth

In this paper, we propose a dense depth estimation pipeline for multiview 360∘ images. The proposed leverages spherical camera model that compensates radial distortion in key contribution of paper is the extension to by introducing translation scaling scheme. Moreover, an effective method setting virtual and minimizing photonic reprojection error. We validate performance using images natural scenes as well synthesized dataset quantitive evaluation. experimental results verify improves accuracy compared current state-of-art methods.